Magne Skålevik Brekke & Strand, Oslo www.akutek.info Room Acoustical Parameters CAN CONCERT HALL PREFERENCE BE PREDICTED? AIA-DAGA meeting, Merano, 19-21 March 2013 What concert hall preference? . Which halls are more preferred, or less preferred, by the average concert-goer? . Why predict preference? Scientifical or academical reasons Curiosity Basis for decisions (Building Committee) RISKY BUSINESS . Null Hypothesis: Preference is purely random 5 aspects 5 parameters ISO-3382* Subjective listener aspect Physical quantity (Parameter) notation and unit Subjective level of sound Sound Strength G (dB) Perceived reverberance Early Decay Time EDT (s) Perceived clarity of sound Clarity C80 (dB) Apparent Source Width Early Lateral Energy LF (1) Fraction Listeners Envelopment Late (*Lateral) Sound Level GL (dB) TVr-predictors (Barron Revised Theory) Difference (in JND units) between TVr-prediction and 126 measurements in 11 halls. 2,0 1,7 1,5 1,5 1,2 1,2 0,9 TVr 1,0 0,7 JND 0,5 0,0 EDT G C LF GL all Does parameter data correlate with subjective preference data? . Let’s take parameter data for 10 of the halls included in Beraneks rank ordering of 58 halls . The 10 halls Musikverein, Vienna Concertgebouw, Amsterdam St David, Cardiff Gasteig, Munich Konserthus, Gøteborg Festspielhaus, Salzburg Liederhalle, Stuttgart Usher, Edinburg Royal Festival Hall, London Barbican, London 10 halls: Objective vs Subjective Class I II III EDT G C LF Glate G125 Criterion 2.0 4.3 -0.3 0.20 1.5 4.7 Tolerance ± 1.1 1.0 0.7 1.1 1.0 1.0 10 halls too few? Now take 53 halls . Beranek subjective ranking of 58 concert halls . Toccupied and geometrical data in 53 of 58 halls . Using TVr-predictors (Barron Revised Theory) . Our set of physical data: T EDT G C GLate G125 W H/W 53 halls out of Beranek’s 58 Rank Rank Vienna Grosser Musikverinsaal 1 Lenox, Tanglewood Music Shed 30 Boston Symphony Hall 2 Munich, Philharmonie Am Gasteig 30 Buenos Airos, Teatro Colon 3 Osaka, Symphony Hall 30 Berlin Konzerthaus (Shauspielhaus) 4 Rotterdam De Doelen 30 Amsterdam Concertgebouw 5 Tokyo, Metropolitan Art Space 30 Tokyo Opera City, Concert Hall 6 Tokyo, Orchard Hall 30 Zurich Grosser Tonhalsaal 7 Toronto, Roy Thompson Hall 30 New York Carnegie Hall 8 Vienna Konzerthaus 30 Basel Stadt Casino 9 Washington, DC, JFK Concert Hall 30 Cardiff, St Davis Hall 10 Salzburg. Festspielhaus 40 Dallas, Meyerson Symphony Center 11 Stuttgart. Liederhalle Grosser Saal 41 Lenox, MA, Seiji Ozawa Hall 13 New York, Avery Fisher Hall 42 Costa Mesa, Segerstrøm Hall 14 Copenhagen Radiohuset Studio 1 43 Salt Lake City, Symphony Hall 15 Edinburgh, Usher Hall 44 Berlin Philharmonie 16 Glasgow, Royal Concert Hall 45 Tokyo, Suntory Hall 17 London Royal Festival Hall 46 Tokyo, Bunka Kaikan 18 Liverpool, Philharmonic Hall 47 Brussels, Palais des Beaux-Arts 19 Paris, Salle Pleyel 49 Baltimore, Meyerhoff Symphony Hall 20 Montreal, Salle Wilifrid-Pelletier 51 Bonn Beethovenhalle 30 Tokyo, NHK Hall 52 Chicago, Orchestra Hall 30 Sydney Opera House Concert Hall 53 Christchurch, Town Hall 30 San Fransisco, Davies Hall 54 Cleveland, Severance Hall 30 Tel Aviv, Frederic Mann Auditorium 55 Gothenburg Concert House 30 London, Barbican Concert Hall 56 Jerusalem, Binyanei Ha'Onnoh 30 Buffalo Kleinhans Music Hall 57 Kyoto Concert Hall 30 London, Royal Albert Hall 58 Leipzig, Gewandhaus 30 Subjective-Objective correlation converge with more data, 53 halls Objective data available: T EDT G C GLate G125 W H/W 70% 60% 50% 40% 30% 20% correlation r^2 correlation 10% 0% 0 1 2 3 4 5 6 7 8 Number of parameters used How does size of random subset affect correlation uncertainty? 100% 10 halls 90% 80% 70% 60% 50% 40% 30% Correlation r^2 20% 10% 0% - 10 20 30 40 50 Selection size Curious scientists ask for correlation, r2=0.?? 100% 80% y = -0,01x + 0,56 criteria 60% R² = 0,63 40% fulfilled 20% Objective Objective 0% 0 10 20 30 40 50 60 Subjective ranking Criteria T EDT G C GLate H/W G125 W Hall average 2.0 s 2.0 s 4 dB 0 dB 3 dB 1.0 4 dB 23 m Tolerance ± 0.1 s 0.2 s 1 dB 1 dB 1 dB 0.2 1 dB 5 m Building committee says: «Halls in Beranek Class II and III are unacceptable» 100% 80% Acceptance limit criteria 60% Class II 40% fulfilled 20% Class III Objective Objective 0% 0 10 20 30 40 50 60 Subjective ranking Criteria T EDT G C GLate H/W G125 W Hall average 2.0 s 2.0 s 4 dB 0 dB 3 dB 1.0 4 dB 23 m Tolerance ± 0.1 s 0.2 s 1 dB 1 dB 1 dB 0.2 1 dB 5 m Consequence of acceptance limit: 8 of 20 Class I halls not good for replication 100% 8 of the Top 20 halls 80% A contradition? criteria 60% Class II 40% fulfilled 20% Class III Objective Objective 0% 0 10 20 30 40 50 60 Subjective ranking Criteria T EDT G C GLate H/W G125 W Hall average 2.0 s 2.0 s 4 dB 0 dB 3 dB 1.0 4 dB 23 m Tolerance ± 0.1 s 0.2 s 1 dB 1 dB 1 dB 0.2 1 dB 5 m Risk: Newer halls (1990+) not included in ranking 100% 80% Are there any BAD halls in this region? criteria 60% 40% fulfilled 20% Objective Objective 0% 0 10 20 30 40 50 60 Subjective ranking Criteria T EDT G C GLate H/W G125 W Hall average 2.0 s 2.0 s 4 dB 0 dB 3 dB 1.0 4 dB 23 m Tolerance ± 0.1 s 0.2 s 1 dB 1 dB 1 dB 0.2 1 dB 5 m Consequences of criteria: These 11 halls among the Top Twenty could be replicated. Rank V N r avr T EDT G C GL H/W W score Vienna Grosser Musikverinsaal 1 15000 1680 29 2,0 2,0 4 -1 1 0,9 20 75 % Berlin Konzerthaus (Shauspielhaus) 4 15000 1575 25 2,0 1,9 4 -1 2 0,9 21 75 % Tokyo Opera City, Concert Hall 6 15300 1636 30 2,0 1,9 4 -1 1 1,1 20 75 % Boston Symphony Hall 2 18750 2625 32 1,9 1,9 2 0 -1 0,8 23 63 % Amsterdam Concertgebouw 5 18780 2037 29 2,0 1,9 3 -1 0 0,6 28 63 % Dallas, Meyerson Symphony Center 11 23900 2065 30 2,8 1,9 4 -3 2 1,0 26 63 % Zurich Grosser Tonhalsaal 7 11400 1546 28 2,1 2,0 5 -1 3 0,7 20 50 % Basel Stadt Casino 9 10500 1448 25 1,8 1,7 5 0 2 0,7 21 50 % Cardiff, St Davis Hall 10 22000 1952 29 2,0 1,9 2 0 -1 0,7 27 50 % Berlin Philharmonie 16 21000 2218 34 1,9 1,8 2 0 -1 0,3 43 50 % Baltimore, Meyerhoff Symphony Hall 20 21530 2467 34 2,0 2,0 2 -1 -1 0,6 29 50 % Criteria T EDT G C GLate H/W G125 W Hall average 2.0 s 2.0 s 4 dB 0 dB 3 dB 1.0 4 dB 23 m Tolerance ± 0.1 s 0.2 s 1 dB 1 dB 1 dB 0.2 1 dB 5 m Consequences of criteria: These 8 Top Twenty halls would NOT be replicated. At what cost? Rank V N r avr T EDT G C GL H/W W score Buenos Airos, Teatro Colon 3 21524 2487 31 1,6 1,6 1 1 -3 1,1 24 38 % New York Carnegie Hall 8 24270 2804 31 1,8 1,7 1 0 -2 0,9 26 38 % Lenox, MA, Seiji Ozawa Hall 13 11610 1180 28 1,7 1,6 4 0 1 0,7 21 38 % Tokyo, Suntory Hall 17 21000 2006 31 2,0 1,9 2 -1 -1 0,5 31 38 % Tokyo, Bunka Kaikan (Ueno) 18 17300 2327 31 1,5 1,5 1 1 -2 0,7 27 38 % Brussels, Palais des Beaux-Arts 19 12520 2150 29 1,6 1,6 3 0 0 1,3 23 38 % Costa Mesa, Segerstrøm Hall 14 27800 2903 37 1,6 1,5 -1 1 -5 0,6 42 13 % Salt Lake City, Symphony Hall 15 19500 2812 34 1,7 1,6 1 0 -2 0,6 30 13 % One cost may be Loss of Freedom. The question is left for discussion Criteria T EDT G C GLate H/W G125 W Hall average 2.0 s 2.0 s 4 dB 0 dB 3 dB 1.0 4 dB 23 m Tolerance ± 0.1 s 0.2 s 1 dB 1 dB 1 dB 0.2 1 dB 5 m Consequences of criteria: These UNRANKED halls could be replicated. Subjective verification required. Rank V N r avr T EDT G C GL H/W W score Manchester Bridgewater Hall - 25000 2357 25 2,0 1,9 2 0 -1 0,9 26 75 % Lucerne, Cultural Ctr. Concert Hall - 18000 1892 23 2,1 2,0 4 -1 1 1,1 22 75 % Valencia, Paleu de la Musica - 15400 1790 18 2,1 2,0 5 -1 2 0,8 21 75 % Taipei Cultural Centre, Concert Hall - 16700 2074 26 2,0 2,0 4 -1 1 0,7 27 63 % Fort Worth, Bass Performance Hall - 27300 2072 26 2,0 1,9 2 0 -1 1,2 22 63 % Hong Kong, Cul.